War Deaths by Nasir Sharaf
Steven Pinker has argued in his best selling book “the Better Angels of Our Nature” that human violence has been trending downward. I wanted to validate his claims, so I used one of his datasets - the Uppsala Conflict Data Program Battle related Deaths Dataset. The exploration is as follows:
## conflict_id dyad_id year
## 1 13349 14273 2013
## 2 13349 14273 2014
## 3 13349 14273 2015
## 4 13721 14847 2015
## 5 13692 11199 2001
## 6 13641 14669 2016
## location_inc
## 1 Myanmar (Burma)
## 2 Myanmar (Burma)
## 3 Myanmar (Burma)
## 4 Algeria
## 5 Afghanistan, United Kingdom, United States of America
## 6 Nigeria
## side_a
## 1 Government of Myanmar (Burma)
## 2 Government of Myanmar (Burma)
## 3 Government of Myanmar (Burma)
## 4 Government of Algeria
## 5 Government of Afghanistan
## 6 Government of Nigeria
## side_a_2nd
## 1
## 2
## 3
## 4
## 5
## 6 Government of Cameroon, Government of Chad, Government of Niger
## side_b
## 1 PSLF
## 2 PSLF
## 3 PSLF
## 4 IS
## 5 Government of United Kingdom, Government of United States of America
## 6 IS
## side_b_id sideb2nd incompatibility territory_name bd_best bd_low bd_high
## 1 5965 1 Palaung 29 29 30
## 2 5965 1 Palaung 37 37 203
## 3 5965 1 Palaung 106 106 181
## 4 234 1 Islamic State 26 26 26
## 5 28, 3 2 1397 1388 2075
## 6 234 1 Islamic State 2213 2210 2311
## type_of_conflict battle_location gwno_a gwno_a_2nd gwno_b gwno_b_2nd
## 1 3 Myanmar (Burma) 775
## 2 3 Myanmar (Burma) 775
## 3 3 Myanmar (Burma) 775
## 4 3 Algeria 615
## 5 2 Afghanistan 700 200, 2
## 6 4 Nigeria 475 471, 483, 436
## gwno_loc gwno_battle region version
## 1 775 775 3 17
## 2 775 775 3 17
## 3 775 775 3 17
## 4 615 615 4 17
## 5 2, 200, 700 700 1, 3, 5 17
## 6 475 475 4 17Univariate Plots Section
## [1] "conflict_id" "dyad_id" "year"
## [4] "location_inc" "side_a" "side_a_2nd"
## [7] "side_b" "side_b_id" "sideb2nd"
## [10] "incompatibility" "territory_name" "bd_best"
## [13] "bd_low" "bd_high" "type_of_conflict"
## [16] "battle_location" "gwno_a" "gwno_a_2nd"
## [19] "gwno_b" "gwno_b_2nd" "gwno_loc"
## [22] "gwno_battle" "region" "version"I first just wanted to look at the dataset and see what variables we have to work with. Some of the variables are confusing, so I consulted the codebook to better understand that dataset.
There were two datasets, one that was a “Dyadic” set, a collection of two-party battles, and the set I used which was a conflict set. The Dyadic ID was carried over from the dyadic set.
Incompatibility returns 1 or 2. 1 stands for ‘incompatibility about government’ indicating that the source of the conflict is a discrepency over governmental affairs. 2 stands for ‘incompatibility about territory’ indicating a resource based conflict.
Another prominent variable will be the type_of_conflict variable. It is also given a numeric evaluation from 1 to 4. 1 stands for extrasystemic - I will be honest, the codebook is lacking for explanatinon as to what this could mean. 2 stands for interstate, meaning two or more countries at war with each other. 3 stands for internal, so mostly civil war or revolutions that do not involve other state entities. 4 stands for international internal, which would refer to a conflict in which a state is deal with internal unrest but has also accepted support from foreign troops.
The first real plot was to set battle deaths against the years they occured. I then filtered by type of conflict, to see if battle deaths were overrepresented by one type of conflict. I could see that battle deaths by raw count were far, far more prevalent in internal conflicts (type 3) but internationalized internal conflicts (type 4), that is internal conflicts in which foreign troops were present, were rising in recent years.
Type of conflict tells us who the main conflict is with, but not what the conflict is about. I thought it would be best to see battle deaths by incompatibility. The codebook tells us that there is only two categories: 1 stands for ‘incompatibility about government’ indicating that the source of the conflict is a discrepency over governmental affairs. 2 stands for ‘incompatibility about territory’ indicating a resource based conflict. However, there was a “3” that was plotted, so this had to be removed. We can see that there seems to be an even split between type 1 incompatibility and type 2 incompatibility.
Type of conflict also doesn’t tell us where the conflict is, so the next area to explore was the region of battle deaths. There was a similar moment of data cleaning to clean out some battles that spanned a few regions. We can see that region three (asia) and region four (africa) had the highest battle deaths. Region two (the middle east) seems to be trending up while region five (the Americas) is trending down.
## bd$battle_location: Afghanistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30 1975 3512 3651 5189 8937
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Afghanistan, Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 4734 5732 6705 6522 7326 8048
## --------------------------------------------------------
## bd$battle_location: Afghanistan, India
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 281 281 281 281 281 281
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 54.0 752.8 2276.0 5227.4 7506.8 16581.0
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, Saudi Arabia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 108 177 246 244 312 378
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, Saudi Arabia, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 174 174 174 174 174 174
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 38.0 161.8 214.0 229.0 281.2 450.0
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, Somalia, Yemen (North Yemen)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 290 290 290 290 290 290
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 65.0 74.5 84.0 84.0 93.5 103.0
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Pakistan, United States of America
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1583 1583 1583 1583 1583 1583
## --------------------------------------------------------
## bd$battle_location: Afghanistan, Tajikistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 368.0 624.8 881.5 881.5 1138.2 1395.0
## --------------------------------------------------------
## bd$battle_location: Albania, Serbia (Yugoslavia)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1404 1404 1404 1404 1404 1404
## --------------------------------------------------------
## bd$battle_location: Algeria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.0 236.0 472.0 724.3 931.0 3029.0
## --------------------------------------------------------
## bd$battle_location: Algeria, Chad, Niger
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 451 451 451 451 451 451
## --------------------------------------------------------
## bd$battle_location: Algeria, Mali
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 229 229 229 229 229 229
## --------------------------------------------------------
## bd$battle_location: Angola
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25 37 72 1350 1123 12054
## --------------------------------------------------------
## bd$battle_location: Angola, DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3924 3924 3924 3924 3924 3924
## --------------------------------------------------------
## bd$battle_location: Angola, Namibia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 461.0 730.5 1000.0 1244.7 1636.5 2273.0
## --------------------------------------------------------
## bd$battle_location: Argentina, Israel, Lebanon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 87 87 87 87 87 87
## --------------------------------------------------------
## bd$battle_location: Armenia, Azerbaijan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 48.5 141.0 729.7 1339.0 2167.0
## --------------------------------------------------------
## bd$battle_location: Azerbaijan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.00 32.00 35.00 40.14 45.00 63.00
## --------------------------------------------------------
## bd$battle_location: Bangladesh
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.00 29.75 43.00 42.67 54.00 62.00
## --------------------------------------------------------
## bd$battle_location: Belgium, Iraq, Libya, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 11846 11846 11846 11846 11846 11846
## --------------------------------------------------------
## bd$battle_location: Bhutan, India
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 34 34 34 34 34 34
## --------------------------------------------------------
## bd$battle_location: Bosnia-Herzegovina
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 133 526 1197 1332 2053 2835
## --------------------------------------------------------
## bd$battle_location: Bosnia-Herzegovina, Bosnia-Herzegovina, Croatia, Serbia (Yugoslavia)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 4018 4018 4018 4018 4018 4018
## --------------------------------------------------------
## bd$battle_location: Bosnia-Herzegovina, Croatia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 116.0 301.5 487.0 487.0 672.5 858.0
## --------------------------------------------------------
## bd$battle_location: Bosnia-Herzegovina, Croatia, Croatia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 126 126 126 126 126 126
## --------------------------------------------------------
## bd$battle_location: Burundi
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 49.0 135.0 279.0 428.7 598.5 1348.0
## --------------------------------------------------------
## bd$battle_location: Burundi, Burundi, DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1114 1114 1114 1114 1114 1114
## --------------------------------------------------------
## bd$battle_location: Cambodia (Kampuchea)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 118.0 247.0 274.0 404.6 579.0 788.0
## --------------------------------------------------------
## bd$battle_location: Cambodia (Kampuchea), Cambodia (Kampuchea), Thailand
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 427.0 524.8 622.5 622.5 720.2 818.0
## --------------------------------------------------------
## bd$battle_location: Cambodia (Kampuchea), Thailand
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.0 80.5 132.0 132.0 183.5 235.0
## --------------------------------------------------------
## bd$battle_location: Cameroon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 32.0 44.0 56.0 266.7 384.0 712.0
## --------------------------------------------------------
## bd$battle_location: Cameroon, Nigeria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 189 236 283 1751 2532 4780
## --------------------------------------------------------
## bd$battle_location: Central African Republic
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.0 35.0 45.0 106.4 139.5 325.0
## --------------------------------------------------------
## bd$battle_location: Central African Republic, Chad
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 132 132 132 132 132 132
## --------------------------------------------------------
## bd$battle_location: Central African Republic, DR Congo (Zaire), DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 100 291 482 482 673 864
## --------------------------------------------------------
## bd$battle_location: Central African Republic, DR Congo (Zaire), South Sudan, Sudan, DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 136 136 136 136 136 136
## --------------------------------------------------------
## bd$battle_location: Central African Republic, DR Congo (Zaire), Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 238 238 238 238 238 238
## --------------------------------------------------------
## bd$battle_location: Central African Republic, DR Congo (Zaire), Sudan, DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 120 120 120 120 120 120
## --------------------------------------------------------
## bd$battle_location: Chad
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 34.0 97.0 161.0 360.7 394.0 1250.0
## --------------------------------------------------------
## bd$battle_location: Chad, Chad, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1352 1352 1352 1352 1352 1352
## --------------------------------------------------------
## bd$battle_location: Chad, Chad, Sudan, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 315 315 315 315 315 315
## --------------------------------------------------------
## bd$battle_location: Chad, Niger, Nigeria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 2667 2667 2667 2667 2667 2667
## --------------------------------------------------------
## bd$battle_location: Chad, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 42 42 42 42 42 42
## --------------------------------------------------------
## bd$battle_location: Chad, Sudan, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 353 353 353 353 353 353
## --------------------------------------------------------
## bd$battle_location: China
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 36 36 36 36 36 36
## --------------------------------------------------------
## bd$battle_location: China, Myanmar (Burma)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 80.0 137.2 194.5 194.5 251.8 309.0
## --------------------------------------------------------
## bd$battle_location: Colombia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30.0 217.8 642.5 730.1 1110.2 2263.0
## --------------------------------------------------------
## bd$battle_location: Colombia, Colombia, Ecuador
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 244 244 244 244 244 244
## --------------------------------------------------------
## bd$battle_location: Colombia, Venezuela
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 316 316 316 316 316 316
## --------------------------------------------------------
## bd$battle_location: Comoros
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.00 35.75 42.50 42.50 49.25 56.00
## --------------------------------------------------------
## bd$battle_location: Congo
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 51.0 53.0 167.0 838.8 651.0 3272.0
## --------------------------------------------------------
## bd$battle_location: Congo, Congo, DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 10033 10033 10033 10033 10033 10033
## --------------------------------------------------------
## bd$battle_location: Croatia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 345 345 345 345 345 345
## --------------------------------------------------------
## bd$battle_location: Djibouti
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.00 33.50 37.00 53.33 54.00 129.00
## --------------------------------------------------------
## bd$battle_location: Djibouti, Somalia, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 379.0 961.8 1544.5 1544.5 2127.2 2710.0
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 41.0 122.5 208.0 809.1 905.0 4457.0
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), DR Congo (Zaire), Rwanda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 703 872 1041 1041 1210 1379
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Rwanda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 83.0 359.5 636.0 605.0 866.0 1096.0
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 53 53 53 53 53 53
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Sudan, Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 221 221 221 221 221 221
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 182.0 225.2 268.5 268.5 311.8 355.0
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Uganda, Sudan, Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 189.0 413.2 637.5 637.5 861.8 1086.0
## --------------------------------------------------------
## bd$battle_location: DR Congo (Zaire), Uganda, Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 388.0 610.5 833.0 820.3 1036.5 1240.0
## --------------------------------------------------------
## bd$battle_location: Ecuador, Peru
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 99 99 99 99 99 99
## --------------------------------------------------------
## bd$battle_location: Egypt
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.00 80.25 124.00 203.50 205.00 750.00
## --------------------------------------------------------
## bd$battle_location: Egypt, Ethiopia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 191 191 191 191 191 191
## --------------------------------------------------------
## bd$battle_location: El Salvador
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 88 468 848 1954 2886 4925
## --------------------------------------------------------
## bd$battle_location: Eritrea
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 25.0 28.0 34.5 37.5 57.0
## --------------------------------------------------------
## bd$battle_location: Eritrea, Ethiopia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1000 24096 47192 32731 48596 50000
## --------------------------------------------------------
## bd$battle_location: Ethiopia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 25.0 25.0 1913.3 74.5 30633.0
## --------------------------------------------------------
## bd$battle_location: Ethiopia, Ethiopia, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 211 211 211 211 211 211
## --------------------------------------------------------
## bd$battle_location: Ethiopia, Kenya
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 25.0 25.0 189.5 189.5 683.0
## --------------------------------------------------------
## bd$battle_location: Ethiopia, Kenya, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 940.0 993.8 1047.5 1047.5 1101.2 1155.0
## --------------------------------------------------------
## bd$battle_location: Ethiopia, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 37.5 50.0 67.0 88.0 126.0
## --------------------------------------------------------
## bd$battle_location: Ethiopia, Sudan, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1259 1259 1259 1259 1259 1259
## --------------------------------------------------------
## bd$battle_location: France, Iraq, Libya, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 11557 11557 11557 11557 11557 11557
## --------------------------------------------------------
## bd$battle_location: France, Israel, Lebanon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 33 33 33 33 33 33
## --------------------------------------------------------
## bd$battle_location: France, Sri Lanka
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 3868 3868 3868 3868 3868 3868
## --------------------------------------------------------
## bd$battle_location: Georgia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.00 66.25 143.00 404.25 579.75 1611.00
## --------------------------------------------------------
## bd$battle_location: Georgia, Russia (Soviet Union)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1289 1289 1289 1289 1289 1289
## --------------------------------------------------------
## bd$battle_location: Germany, Netherlands, United Kingdom
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 44 44 44 44 44 44
## --------------------------------------------------------
## bd$battle_location: Germany, United Kingdom
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 38 38 38 38 38 38
## --------------------------------------------------------
## bd$battle_location: Guatemala
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.00 47.50 67.00 60.86 77.00 83.00
## --------------------------------------------------------
## bd$battle_location: Guinea
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 217 217 217 217 217 217
## --------------------------------------------------------
## bd$battle_location: Guinea-Bissau
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 199.0 275.5 352.0 352.0 428.5 505.0
## --------------------------------------------------------
## bd$battle_location: Guinea-Bissau, Senegal
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 47.00 54.75 62.50 62.50 70.25 78.00
## --------------------------------------------------------
## bd$battle_location: Guinea, Liberia, Sierra Leone
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 432 432 432 432 432 432
## --------------------------------------------------------
## bd$battle_location: Haiti
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 48.0 104.0 160.0 138.3 183.5 207.0
## --------------------------------------------------------
## bd$battle_location: India
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.00 35.25 76.00 252.89 264.25 2673.00
## --------------------------------------------------------
## bd$battle_location: India, Myanmar (Burma)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 43 43 43 43 43 43
## --------------------------------------------------------
## bd$battle_location: India, Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 29.5 68.5 462.2 198.2 2353.0
## --------------------------------------------------------
## bd$battle_location: Indonesia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26 51 187 268 286 915
## --------------------------------------------------------
## bd$battle_location: Iran
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 28.00 30.25 52.50 66.50 88.75 133.00
## --------------------------------------------------------
## bd$battle_location: Iran, Iran, Iraq
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 57.00 60.75 64.50 64.50 68.25 72.00
## --------------------------------------------------------
## bd$battle_location: Iran, Iraq
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.0 32.0 42.5 102.1 116.8 356.0
## --------------------------------------------------------
## bd$battle_location: Iran, Iraq, Iran, Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 121 121 121 121 121 121
## --------------------------------------------------------
## bd$battle_location: Iran, Iraq, Italy, Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 76 76 76 76 76 76
## --------------------------------------------------------
## bd$battle_location: Iraq
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25 94 565 1094 1870 4094
## --------------------------------------------------------
## bd$battle_location: Iraq, Kuwait, Philippines, Saudi Arabia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 21790 21790 21790 21790 21790 21790
## --------------------------------------------------------
## bd$battle_location: Iraq, Lebanon, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 68503 68503 68503 68503 68503 68503
## --------------------------------------------------------
## bd$battle_location: Iraq, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 13443 13443 13443 13443 13443 13443
## --------------------------------------------------------
## bd$battle_location: Iraq, Syria, Turkey
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1214 1214 1214 1214 1214 1214
## --------------------------------------------------------
## bd$battle_location: Iraq, Turkey
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30 250 811 1306 1735 4183
## --------------------------------------------------------
## bd$battle_location: Israel
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.0 60.0 120.0 330.3 459.5 1671.0
## --------------------------------------------------------
## bd$battle_location: Israel, Israel, Lebanon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 328 328 328 328 328 328
## --------------------------------------------------------
## bd$battle_location: Israel, Israel, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 379 379 379 379 379 379
## --------------------------------------------------------
## bd$battle_location: Israel, Lebanon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 28.0 41.5 74.0 175.6 109.5 825.0
## --------------------------------------------------------
## bd$battle_location: Israel, Malta
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 34 34 34 34 34 34
## --------------------------------------------------------
## bd$battle_location: Ivory Coast
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 35.0 44.0 93.0 211.2 260.2 624.0
## --------------------------------------------------------
## bd$battle_location: Ivory Coast, Liberia, Liberia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 347 347 347 347 347 347
## --------------------------------------------------------
## bd$battle_location: Jordan, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 34 8517 17000 17000 25484 33967
## --------------------------------------------------------
## bd$battle_location: Kenya, Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 42.0 331.2 1362.0 1233.2 1840.0 2646.0
## --------------------------------------------------------
## bd$battle_location: Kuwait
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1058 1058 1058 1058 1058 1058
## --------------------------------------------------------
## bd$battle_location: Kyrgyzstan, Tajikistan, Uzbekistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 67 67 67 67 67 67
## --------------------------------------------------------
## bd$battle_location: Kyrgyzstan, Uzbekistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 471 471 471 471 471 471
## --------------------------------------------------------
## bd$battle_location: Laos
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30.00 84.75 139.50 139.50 194.25 249.00
## --------------------------------------------------------
## bd$battle_location: Lebanon
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 28.0 45.0 79.0 194.2 115.0 711.0
## --------------------------------------------------------
## bd$battle_location: Lebanon, Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 54547 54547 54547 54547 54547 54547
## --------------------------------------------------------
## bd$battle_location: Lesotho
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 68 68 68 68 68 68
## --------------------------------------------------------
## bd$battle_location: Liberia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 27.0 57.0 341.0 540.8 492.0 1787.0
## --------------------------------------------------------
## bd$battle_location: Liberia, Sierra Leone
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 415 415 415 415 415 415
## --------------------------------------------------------
## bd$battle_location: Libya
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 118.0 156.0 322.0 840.8 1678.0 1930.0
## --------------------------------------------------------
## bd$battle_location: Macedonia, FYR, Serbia (Yugoslavia)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 72 72 72 72 72 72
## --------------------------------------------------------
## bd$battle_location: Malaysia, Philippines
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 70 70 70 70 70 70
## --------------------------------------------------------
## bd$battle_location: Mali
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 34.00 44.25 90.50 87.00 103.50 203.00
## --------------------------------------------------------
## bd$battle_location: Mali, Mali, Niger
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 839 839 839 839 839 839
## --------------------------------------------------------
## bd$battle_location: Mali, Mauritania
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.00 26.75 27.50 27.50 28.25 29.00
## --------------------------------------------------------
## bd$battle_location: Mexico
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 37 64 91 91 118 145
## --------------------------------------------------------
## bd$battle_location: Moldova
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 585 585 585 585 585 585
## --------------------------------------------------------
## bd$battle_location: Morocco
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 207 207 207 207 207 207
## --------------------------------------------------------
## bd$battle_location: Mozambique
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.0 97.5 729.0 735.7 1281.8 1577.0
## --------------------------------------------------------
## bd$battle_location: Myanmar (Burma)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 30.5 68.0 125.3 126.2 1163.0
## --------------------------------------------------------
## bd$battle_location: Myanmar (Burma), Myanmar (Burma), Thailand
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 114.0 115.8 117.5 117.5 119.2 121.0
## --------------------------------------------------------
## bd$battle_location: Myanmar (Burma), Thailand
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 33.0 35.5 117.0 256.9 288.5 1000.0
## --------------------------------------------------------
## bd$battle_location: Nepal
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 40.0 203.0 429.0 901.5 1222.5 3947.0
## --------------------------------------------------------
## bd$battle_location: Nicaragua
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 39.0 160.8 282.5 282.5 404.2 526.0
## --------------------------------------------------------
## bd$battle_location: Niger
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.00 40.00 50.00 87.56 81.00 274.00
## --------------------------------------------------------
## bd$battle_location: Nigeria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 61.0 177.0 405.0 845.1 1629.0 2213.0
## --------------------------------------------------------
## bd$battle_location: Pakistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 38.0 64.0 89.5 1028.1 871.0 6303.0
## --------------------------------------------------------
## bd$battle_location: Panama
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 77.0 268.5 460.0 460.0 651.5 843.0
## --------------------------------------------------------
## bd$battle_location: Papua New Guinea
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.0 29.0 34.0 44.4 55.0 78.0
## --------------------------------------------------------
## bd$battle_location: Papua New Guinea, Solomon Islands
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 35 35 35 35 35 35
## --------------------------------------------------------
## bd$battle_location: Paraguay
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 150 150 150 150 150 150
## --------------------------------------------------------
## bd$battle_location: Peru
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 45.0 93.0 441.7 854.0 1827.0
## --------------------------------------------------------
## bd$battle_location: Philippines
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 32.0 114.5 193.0 299.9 313.0 1514.0
## --------------------------------------------------------
## bd$battle_location: Qatar, Russia (Soviet Union)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1151 1151 1151 1151 1151 1151
## --------------------------------------------------------
## bd$battle_location: Rumania
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 249 249 249 249 249 249
## --------------------------------------------------------
## bd$battle_location: Russia (Soviet Union)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30.0 144.5 290.0 793.4 675.0 5769.0
## --------------------------------------------------------
## bd$battle_location: Russia (Soviet Union), Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 30 30 30 30 30 30
## --------------------------------------------------------
## bd$battle_location: Russia (Soviet Union), Turkey
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 359 359 359 359 359 359
## --------------------------------------------------------
## bd$battle_location: Russia (Soviet Union), Ukraine
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 712 712 712 712 712 712
## --------------------------------------------------------
## bd$battle_location: Rwanda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 87.0 121.5 353.0 701.0 1090.0 2044.0
## --------------------------------------------------------
## bd$battle_location: Saudi Arabia, Yemen (North Yemen)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 2536 2536 2536 2536 2536 2536
## --------------------------------------------------------
## bd$battle_location: Saudi Arabia, Yemen (North Yemen), Yemen (North Yemen)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 6532 6532 6532 6532 6532 6532
## --------------------------------------------------------
## bd$battle_location: Senegal
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25.0 37.5 145.5 156.0 256.2 319.0
## --------------------------------------------------------
## bd$battle_location: Serbia (Yugoslavia)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 64.0 100.0 673.5 1092.2 1665.8 2958.0
## --------------------------------------------------------
## bd$battle_location: Sierra Leone
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 48.0 424.5 599.0 1105.8 1291.0 3287.0
## --------------------------------------------------------
## bd$battle_location: Somalia
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25 228 547 1362 1491 8005
## --------------------------------------------------------
## bd$battle_location: South Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 137.0 285.2 613.0 695.3 877.8 1667.0
## --------------------------------------------------------
## bd$battle_location: South Sudan, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 367 367 367 367 367 367
## --------------------------------------------------------
## bd$battle_location: South Sudan, Sudan, Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 856 958 1060 1060 1162 1264
## --------------------------------------------------------
## bd$battle_location: Spain
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 44 44 44 44 44 44
## --------------------------------------------------------
## bd$battle_location: Sri Lanka
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.0 589.8 2284.5 2868.3 3844.2 10165.0
## --------------------------------------------------------
## bd$battle_location: Sudan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 145 684 1128 1610 2222 4891
## --------------------------------------------------------
## bd$battle_location: Sudan, Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 295.0 668.5 1317.0 1878.1 2931.5 4335.0
## --------------------------------------------------------
## bd$battle_location: Sudan, Uganda, Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 888 888 888 888 888 888
## --------------------------------------------------------
## bd$battle_location: Syria
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 25 56 964 6052 11161 27190
## --------------------------------------------------------
## bd$battle_location: Syria, Turkey
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26 9640 19253 19253 28866 38480
## --------------------------------------------------------
## bd$battle_location: Tajikistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 28.00 76.75 227.00 922.62 1147.50 3376.00
## --------------------------------------------------------
## bd$battle_location: Thailand
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.00 69.75 134.50 127.14 182.75 212.00
## --------------------------------------------------------
## bd$battle_location: Trinidad and Tobago
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 39 39 39 39 39 39
## --------------------------------------------------------
## bd$battle_location: Turkey
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 29.00 77.25 218.50 352.67 310.50 2117.00
## --------------------------------------------------------
## bd$battle_location: Uganda
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 72.0 105.5 385.0 467.5 795.0 1019.0
## --------------------------------------------------------
## bd$battle_location: Ukraine
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 88 255 1304 1040 1558 1996
## --------------------------------------------------------
## bd$battle_location: United Kingdom
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 26.00 26.75 27.50 27.50 28.25 29.00
## --------------------------------------------------------
## bd$battle_location: Uzbekistan
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 37 37 37 37 37 37
## --------------------------------------------------------
## bd$battle_location: Venezuela
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 144 144 144 144 144 144
## --------------------------------------------------------
## bd$battle_location: Yemen (North Yemen)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 94.0 228.2 861.0 964.5 1531.8 2330.0
I also wanted to see a finer breakdown of the data by territory, so I printed the summary. I also attempted a bar plot of the battle deaths per battle location. The codebook explicitly calls out against using battle locations for geospatial analysis, but I only wanted to see which locations have the most battles.
Reflections: Univariate
Although this section has only been devoted to analyzing single variables at a time, we have already learned quite a bit. We have seen regional differences in conflict and how bloody some types of conflict are compared to others. We will continue by comparing two variables agaisnt each other for a deeper understanding of global violence.
Univariate Analysis
What is the structure of your dataset?
I will repeat here what I wrote above when explaining the structure of the data set:
The dataset has an accompanying codebook that illuminates the more confusing variables.
There were two datasets, one that was a “Dyadic” set, a collection of two-party battles, and the set I used which was a conflict set. The Dyadic ID was carried over from the dyadic set.
Incompatibility returns 1 or 2. 1 stands for ‘incompatibility about government’ indicating that the source of the conflict is a discrepency over governmental affairs. 2 stands for ‘incompatibility about territory’ indicating a resource based conflict.
Another prominent variable will be the type_of_conflict variable. It is also given a numeric evaluation from 1 to 4. 1 stands for extrasystemic - I will be honest, the codebook is lacking for explanatinon as to what this could mean. 2 stands for interstate, meaning two or more countries at war with each other. 3 stands for internal, so mostly civil war or revolutions that do not involve other state entities. 4 stands for international internal, which would refer to a conflict in which a state is deal with internal unrest but has also accepted support from foreign troops.
What is/are the main feature(s) of interest in your dataset?
I want to see if the conflicts and battle related deaths have decreaseed or increased over time, as was studied by Steven Pinker. Additionally, if these death increased/decreased across all categories uniformly or whether they decreased or increased in certain types of conflict only.
What other features in the dataset do you think will help support your
investigation into your feature(s) of interest?
The regions may be of interest. Particularly with the incompatibility about territory types of conflict, it would unveil which regions are most prone to resource based conflicts. It would’ve been interesting to compare with population data also, but that may be foregone.
Did you create any new variables from existing variables in the dataset?
No. Pinker has claimed that he used as raw and unfiltered data as possible and I strive to do the same.
Of the features you investigated, were there any unusual distributions?
Did you perform any operations on the data to tidy, adjust, or change the form
of the data? If so, why did you do this?
Yes. There was a 3 given for incompatibility, which should not be there according to the codebook, so that too was dropped. One unusual distribution was the rise in battle deaths count for international internal, which I believe has to do with the Syrian conflict.
Bivariate Plots Section
At first, I just wanted to see battle deaths by year for each battle. This might’ve been better placed in the univariate section, but I also overlayed a max trendline. I plotted one with raw data, one with a square root axis, and one with a log axis. For all except the log axis plot I plotted the line for the maximum battle deaths. For the log plot I plotted the mean line for battle deaths. This sets up what to expect ot of the next series of plots.
Following from the previosu plots, I also wanted to parse out battle deaths by type of conflict (the first plot), incompatibility (the second) and region (the third plot). We saw a rise in type three and type 4 conflicts, civil wars and internationalized civil wars, which I attribute by timeline ot be the syrian conflict. By incompatibility we also see a start rise in incompatibility over resources (2). And of course, when we look at battle deaths by region we can see a sharp spike in region 2, the middle east.
I thought it might be best to use box plots to get a better understanding of the data. We could see with most of the above plots that there were few high casualty battle deaths but many low casualty deaths, which would make a box plot a better choice. I fumbled for a scale, but the log scale seems like the best one. I started, at random, by looking at a box plot for battle deaths by region. Region 2, the middle east, had much higher outliers but the medians seemed to be on par with other regions. In fact, the only region with a lower median was region 3, Asia, althought they had a higher frequency of outliers.
I did a similar thing for incompatibility and battle deaths. As we’ve seen a few times now, incompatibility over territory tends to draw more battle deaths than incompatibility over government.
## # A tibble: 6 x 4
## year bd_best_mean bd_best_med n
## <int> <dbl> <dbl> <int>
## 1 1989 1358. 244. 40
## 2 1990 1626. 203 49
## 3 1991 1351. 156. 52
## 4 1992 744. 165 49
## 5 1993 856. 154 43
## 6 1994 693. 147 48
For this section I grouped the battle deaths by year and cycled through different axises. The mean battle deaths by year with a mean trendline is important, it shows a surprising slight cyclical trend in battle deaths. Its clear that by raw numbers there have been higher battle deaths in the past few years than previously. It also looks like high-casualty battles happen every ten to twelve years (but this maybe the relatively small timeframe over which the data is collected.)
I took the same data set grouped by year and filtered for type of conflict. We’ve already seen to some extent how much more prevalent type 3, civil conflict, and type 4, internationalized civil conflict, is but this plot brought home exactly how much more prevalent civil conflicts are as opposed to interstate conflicts. This is surprising to me because I tend to think of battle deaths as between two or more nations, but it appears that it is more common for battle deaths to occur within states. While interstate conflict and battle deaths are very devastating when they happen, intrastate conflict is much more prevalent.
Looking at battle deaths by year by incompatibility, we can see that type 1 incompatibilty, incompatibility over government, seemed to peak in the late 1990s and by 2000. Conversely, battle deaths by type 2 incompatibilty (over territory) saw a swell at around 2013-2014. Overall though, incompatibilites about government seem to on the whole be declining even though the most prevalent type of conflict, as we saw earlier, was intrastate. This is strange, I would expect most intrastate conflict to be about government/how things are run but it may be pointing to how states are not providing enough territory or resources for their people.
Looking at the log graph again, we seem some familiar ideas playing out. Region 2, the middle east, has a tumultuous ascendency that’s peaked in recent years. That was expected. Region three, Asia, has high frequencies of battle deaths that is probably due to the hgih population and size of the continent. The average of battle deaths in Asia is on par with mos tof the other regions. The surprising region was region 5, the Americas, which has a noticable downward trend.
Reflection: Bivariates
Many of the observations made in the univariate section were expanded upon in this section. For example, we knew from the unvariate analysis that type two incompatibility was responsible for a higher battle death count than type one, but we could see specific peaks and valleys. It seemed like government incompatibilities, while devastting when they occured, were declining compared to territory incompatibilities. Insights like this were made through out this section.
The analysis itself was more nuanced in this section. Many times, the same data had to be plotted against different axises in order to see clear trends or better evaluate the data.
Bivariate Analysis
Talk about some of the relationships you observed in this part of the
investigation. How did the feature(s) of interest vary with other features in
the dataset?
The absolute number of battle deaths seems to have gone up, but does so periodically. I was very surprised at the seeming sinusoidal behavior of it.
Did you observe any interesting relationships between the other features
(not the main feature(s) of interest)?
By incompatibility type we see that conflict with/because of government had a spike in 2000 but conflict due to territory increased around 2015. Looking at conflict type, we see a similar spike for 2 - interstate at around 2000 and a slight bump for conflict type 3 - internal. Not that internal has a low average but is of much higher frequency.
What was the strongest relationship you found?
By type of cnflict, 3 - internal conflict is incredibly frequent. It doesn’t show as much in terms of averages because the casualties are low but they occur far, far more than other data.
Multivariate Plots Section
## # A tibble: 6 x 5
## # Groups: year [2]
## year region mean_bd_best2 median_bd_best2t n
## <int> <fct> <dbl> <dbl> <int>
## 1 1989 1 144. 144. 2
## 2 1989 2 260. 140. 4
## 3 1989 3 747. 194 13
## 4 1989 4 2879. 699 12
## 5 1989 5 968. 238 9
## 6 1990 1 78.7 49 3
The first multivariate graph was plotting the multiple regions, by battle death by year, on one graph. The first graph was using the means of battle deaths and showed about what I expected given the bivariate analysis. There was, as expected, a sharp rise in recent years in region 2, the middle east. There were shapr bumps in region four, Africa, and region one, Europe. THe second grpah, by medians, was very surprising however. This showed that the late 90s and early 00s had the highest median battle deaths in region one, Europe. This might’ve been the Bosnian War or possibly the Irish Peace Process. Either way, this was incredibly surprising since it did not show up in any of the previous analyses.
## # A tibble: 6 x 5
## # Groups: year [3]
## year incompatibility mean_bd_best2 median_bd_best2t n
## <int> <int> <dbl> <dbl> <int>
## 1 1989 1 601. 115 15
## 2 1989 2 1812. 526 25
## 3 1990 1 1517. 76 26
## 4 1990 2 1750. 700 23
## 5 1991 1 1794 100 25
## 6 1991 2 940. 220 27
Showing both forms of incompatibility on the smae graph reiterates exactly the sentiments I believed earlier. Type 2 incompatibility, over territory, is far more prevalent and more responsible for a higher battle dath count than type one. The first graph shows how conflicts over territory have increased to a startling amount in recent years. THe second graph shows the median mbattle deaths by incompatibility and shows that the median deaths are always higher except for one year in 1993.
## # A tibble: 6 x 6
## # Groups: year [2]
## year type_of_conflict sum_bd_best2 mean_bd_best2 median_bd_best2t n
## <int> <fct> <int> <dbl> <dbl> <int>
## 1 1989 2 868 434 434 2
## 2 1989 3 42775 1296. 207 33
## 3 1989 4 10662 2132. 1149 5
## 4 1990 2 1083 542. 542. 2
## 5 1990 3 75195 1709. 142 44
## 6 1990 4 3403 1134. 1126 3
Analyzing by type of conflict shows that, as has been a recurring observation, type 3 conflicts and type four conflicts have been steadily rising globally. This is most prevalent in thefirst graph, which orders by sums. When we look by medians, we can see spiked at around 1990 and 2000 and see that the median battle deaths are highest for type 2, interstate conflicts. Same is true for the plot by mean. Another interesting trend I noticed was the sharp rise at around 2013 in conflicts that were categorized as both civil and internationalized civil.
Reflections: Multivariates
Much of the analysis performed her was a continuation of the anlysis performed in the bivariate section, but was clenaer and produced more insightful graphs and visualizations. It is easier to compare trends against each other on the same plot and not across plots, so that intrinsically revealed some interesting insights.
Multivariate Analysis
Talk about some of the relationships you observed in this part of the
investigation. Were there features that strengthened each other in terms of
looking at your feature(s) of interest?
Type of conflict, when plotted directly against other types of conflict still showed that 2 - inter state conflict caused the most battle deaths at the year 2000. This supports an earlier observation. I also saw a rise in 4 - international internal conflicts and on the region graph a massive spike in region 2 - the middle east after 2010, somewhat supporting my hypothesis of Syria’s effect on the data.
Were there any interesting or surprising interactions between features?
Because of the previous observation of the frequency of type 3 conflicts - internal, I summed by conflict type. Interestingly, it seems as though internal conflicts (3) cycle off with state conflicts (2). Also, as international internal deaths rose, there was a steep drop off in type 3 - internal deaths. Was this a change in how the data was categorized?
OPTIONAL: Did you create any models with your dataset? Discuss the
strengths and limitations of your model.
No, but the cyclical nature seems to indicate a model might exist.
Final Plots and Summary
Plot One
Description One
THe plots show the number of battle deaths by conflict type, where each conflict type was described in the accompanyng codebook as: (1) - extrasystemic, (2) Interstate, (3) Internal (involving no foreign agents) and (4) internationalized internal (internal conflicts with interstate forces). These plots show a few thing well, namely the frequency of internal conflicts and the intermittency of interstate conflicts. While the average for internal conflicts is low its frequency overshadows other types of conflict.
Plot Two
Description Two
These plots show battle deaths by region, where the codebook again reveals that (1) Europe, (2) is the Middle East, (3) is Asia, (4) is Africa and (5) is the Americas. This shows, interestingly, that the violence in the Middle East is of late unprecedented. Also this shows, though hard to see, a slow but steady growth of battle deaths in Asia. This also shows how different violence is distributed regionally.
Plot Three
Description Three
I wanted to show a simple graph, as Pinker would show, of just the average battle deaths per year against the year. It shows, most interestingly a weak sort of sine wave. While alot of these are explained by violence in the middle east or the tragedy of Darfur, its seeming periodicity is striking.
Reflection
Pinker did do one thing to his datasets that I simply was unable to do and that was he made all data shown as per capita. I could not find a compatible enough dataset without truncating too much data, but this is one area for future study. My biggest criticism of Pinker’s analysis has been how he does not compare battle deaths to death rates per population, so this is another area I would like to study.
All in all, I was surprised by the periodicity of the data sometimes. It might’ve been naive analysis, but they do say history repeats itself, so it would make sense that it shows up in the data.